Hi,

Need some opinion on Jitter.

LCaudio XO3 jitter is around 5ppm. I am not sure if that is the same as 5ps.
My DAC's jitter, as per Stereophile is around 174ps.

The sound heard thru the player is inferior as compared to hearing via the
DAC. However, the sound heard thru DIP (Reclocking and jitter reducer) then
to the DAC is much better than without the DIP.

The question is,

If the DAC jitter is already around 174ps how could an external reclocking
device (DIP) reduce that?

Moreover the transport clock's jitter is around 5ps unless others component
add up the jitter such as the cable.

Can any shed some light, pls.

"Chelvam" wrote in message

Hi,

Need some opinion on Jitter.

Three words: AES preprint 4826.

More specifically:

AES Preprint 4826 discusses this on page 23. It says that the threshold of
audibility with a 20 kHz tone is 10 nanoseconds rms, and at 4 kHz the
threshold of audiblity is 100 nanaseconds rms.

However, at 4 kHz, the distortion products from 100 nanoseconds of jitter
are about one fifth the
amplitude as they are with 20 KHz and 100 nanoseconds of jitter.

IOW, the distortion products from 4 KHz and 100 ns of jitter are about the
same size as the ones from 20 KHz and 20 ns of jitter. This accounts for
about half of the observed diffference in perceptual threshold.

We can see that the perceptiion of jitter is approximately controlled by the
size of the garbage that the jitter creates, as it is reduced by masking in
the ear. The paper mentions masking by the musical program material.

Extrapolating this to lower frequencies suggests that at 20 Hz, 20,000 ns (2
uS) is the threshold of audibility for jitter.

We can conclude that the worst case is for musical content at 20 KHz, and at
that point the threshold of audibility is about 20 nanaseconds or 20,000 ps.

LCaudio XO3 jitter is around 5ppm.

The natural basis for a jitter spec in ppm is the clock period, If the clock
frequency is 44 KHz, then the natural basis for a ppm-based spec is about 22
microseconds. 5 times 10 to the minus 6, times 22 times 10 to the minus 6 is
110 times 10 to the minus twelfth. In 1/1000's the sequence goes seconds,
milliseconds, microseconds, nanoseconds, picoseconds. 5 ppm at 44 KHz is
something like 110 picoseconds (ps).

I am not sure if that is the same as 5ps.

Calculations equate that with approximately 110 picseconds (ps).

My DAC's jitter, as per Stereophile is around 174ps.

Per AES preprint 4826, the audibility of jitter varies with the frequency of
the signal being modulated by the jitter. But, their audibility figures are
given in nanoseconds which are 1000 times greater than picoseconds. The
paper suggests strongly that the ear is most sensitive to jitter as applied
to 20 KHz signals. At this point the threshold of audiblity is about 20,000
ps which is far less than 110 ps. IOW, 110 ps jitter is nit.

The sound heard thru the player is inferior as compared to hearing
via the DAC. However, the sound heard thru DIP (Reclocking and jitter
reducer) then to the DAC is much better than without the DIP.

The question is,

If the DAC jitter is already around 174ps how could an external
reclocking device (DIP) reduce that?

Let's look at this situation critically.

My statement of the situation:

A sighted, non-time-synched, non-level-matched listening *test* wth no
other on-site technical tests backing it, is positive for audible
differences.

Somehow, the conclusion is reached that *any* and *all* perceived
differences are due to differences in jitter specs that in fact, may or not
be representative of the equipment being compared, AND is probably inaudible
by a factor of 100 or more.

What's wrong with this picture, logically-speaking?

With all due respect I'd say, just about everything is wrong with this
picture.

Moreover the transport clock's jitter is around 5ps unless others
component add up the jitter such as the cable.

Can any shed some light, pls.

With all due respect I'd say that the listening tests described have about
6 times more uncontrolled variables than it would take to make assigning any
technical cause to its outcome a pure and pristine, nearly science-free
wild-ass guess.

"Arny Krueger" wrote in message



Correction:

Per AES preprint 4826, the audibility of jitter varies with the
frequency of the signal being modulated by the jitter. But, their
audibility figures are given in nanoseconds which are 1000 times
greater than picoseconds. The paper suggests strongly that the ear is
most sensitive to jitter as applied to 20 KHz signals. At this point
the threshold of audiblity is about 20,000 ps which is far greater than
110 ps. IOW, 110 ps jitter is nit.

"Arny Krueger" wrote in message ...
"Arny Krueger" wrote in message



Correction:

Per AES preprint 4826, the audibility of jitter varies with the
frequency of the signal being modulated by the jitter. But, their
audibility figures are given in nanoseconds which are 1000 times
greater than picoseconds. The paper suggests strongly that the ear is
most sensitive to jitter as applied to 20 KHz signals. At this point
the threshold of audiblity is about 20,000 ps which is far greater than
110 ps. IOW, 110 ps jitter is nit.

Thanks Arny, I noticed that. I was trying to figure out if at all
jitter was an issue. And , I do agree with you that under DBT I can't
tell the difference in so many things. In fact, at one point I
downgraded myself from High End to Mid End and was enjoying the music
for about a year till last month………..

Okay, now let's ignore about the audible difference and confine to
technical issues because I believe there are some discrepancies in
jitter buster claims.

The measurement of jitter for DAC "X" was between 174ps to 300++ps
between different signals. Can I safely say that the best measurement
the DAC "X" capable of should be 174ps (or slightly less) from a
"perfect" 100% jitter free input?

In that case, let's say a transport with jitter of 20ns clock sends
digital signal without being subjected to any additional jitter (I
know that is impossible) the total DAC's jitter should be around 174ps
+20ps or slightly more. Correct?

We all know clock's jitter of 20ns is almost the best under current
technology and therefore any jitter buster and reclocking device which
is connected before the DAC could only give the best of around 20ps
[or (8ps;3 sigma) – I don't know what's that.- but that is according
some clock's specification].

So the only advantage of using a jitter buster/reducer device is to
benefit a further reduction of maybe 10ps which is too small for any
audible difference. Correct?

However, jitter buster/reducer device also boost the signal. So is it
possible that it is not the jitter but the boosting of the signal
which makes the audible difference?

Thanks.

"TChelvam" wrote in message
om
"Arny Krueger" wrote in message
...
"Arny Krueger" wrote in message



Correction:

Per AES preprint 4826, the audibility of jitter varies with the
frequency of the signal being modulated by the jitter. But, their
audibility figures are given in nanoseconds which are 1000 times
greater than picoseconds. The paper suggests strongly that the ear
is most sensitive to jitter as applied to 20 KHz signals. At this
point the threshold of audiblity is about 20,000 ps which is far
greater than 110 ps. IOW, 110 ps jitter is nit.

Thanks Arny, I noticed that. I was trying to figure out if at all
jitter was an issue. And , I do agree with you that under DBT I can't
tell the difference in so many things. In fact, at one point I
downgraded myself from High End to Mid End and was enjoying the music
for about a year till last month.....

Okay, now let's ignore about the audible difference and confine to
technical issues because I believe there are some discrepancies in
jitter buster claims.

The measurement of jitter for DAC "X" was between 174ps to 300++ps
between different signals. Can I safely say that the best measurement
the DAC "X" capable of should be 174ps (or slightly less) from a
"perfect" 100% jitter free input?

In that case, let's say a transport with jitter of 20ns clock sends
digital signal without being subjected to any additional jitter (I
know that is impossible) the total DAC's jitter should be around 174ps
+20ps or slightly more. Correct?

We all know clock's jitter of 20ns is almost the best under current
technology and therefore any jitter buster and reclocking device which
is connected before the DAC could only give the best of around 20ps
[or (8ps;3 sigma) - I don't know what's that.- but that is according
some clock's specification].

So the only advantage of using a jitter buster/reducer device is to
benefit a further reduction of maybe 10ps which is too small for any
audible difference. Correct?

However, jitter buster/reducer device also boost the signal. So is it
possible that it is not the jitter but the boosting of the signal
which makes the audible difference?

I really have no idea. This is all too speculative.